Cytomegalovirus is the leading infectious cause of birth defects which can result in deafness and mental retardation in neonates, and can cause severe viral pneumonia and colitis in transplant recipients and sight-threatening retinitis in patients with AIDS. Epstein-Barr virus (EBV) causes infectious mononucleosis and is associated with a number of cancers including Burkitt lymphoma, nasopharyngeal carcinoma, Hodgkin lymphoma, and post-transplant lymphoproliferative disease. Human CMV and EBV naturally infect humans, but not small animals or nonhuman primates. The best models currently available for CMV and EBV are rhesus monkey CMV and EBV. The goal of this study is to develop an effective vaccine for these rhesus viruses and to use these as a model for vaccines for their human counterparts. We are using various approaches including soluble recombinant proteins, recombinant virus vectors expressing viral proteins, and replication defective viruses as vaccines. We initiated construction of a candidate vaccine virus for rhesus CMV. In order to develop a rhesus CMV vaccine virus that would express nearly all of the viral proteins important for inducing a potent immune response, yet be safe in animals, we deleted a viral protein from the virus that is essential for virus growth. The resulting replication-defective virus lacks one viral protein, glycoprotein L (gL), and replicated only in cells expressing rhesus CMV gL. Noncomplementing cells infected with the replication-defective rhesus CMV released intact, noninfectious rhesus CMV particles that were indistinguishable from wild-type virus by electron microscopy. In addition, noncomplementing cells infected with the replication-defective rhesus CMV produced glycoprotein B, the major target of neutralizing antibodies, at levels similar to those observed in cells infected with wild-type virus. We also deleted gL from human CMV and showed that this virus could only grow in cells expressing human CMV gL. We also found that human CMV gL could not complement the replication of rhesus CMV with gL deleted, and that rhesus CMV gL could not complement the replication of human CMV with gL deleted. The human CMV with gL deleted might be a promising vaccine candidate.